The Role of the Rhomboid Superfamily in ER Protein Quality Control: From Mechanisms and Functions to Diseases

Abstract

The endoplasmic reticulum (ER) is an essential organelle in eukaryotic cells and is a major site for protein folding, modification, and lipid synthesis. Perturbations within the ER, such as protein misfolding and high demand for protein folding, lead to dysregulation of the ER protein quality control network and ER stress. Recently, the rhomboid superfamily has emerged as a critical player in ER protein quality control because it has diverse cellular functions, including ER-associated degradation (ERAD), endosome Golgi-associated degradation (EGAD), and ER preemptive quality control (ERpQC). This breadth of function both illustrates the importance of the rhomboid superfamily in health and diseases and emphasizes the necessity of understanding their mechanisms of action. Because dysregulation of rhomboid proteins has been implicated in various diseases, such as neurological disorders and cancers, they represent promising potential therapeutic drug targets. This review provides a comprehensive account of the various roles of rhomboid proteins in the context of ER protein quality control and discusses their significance in health and disease.

The endoplasmic reticulum (ER) is a cellular compartment that is important for the synthesis of lipids, secretory proteins, and transmembrane proteins. Approximately one-third of eukaryotic proteins are cotranslationally imported into the ER, where they are promptly folded and then delivered to various locations for secretion (Hegde and Keenan 2022). Numerous chaperones and enzymes residing in the ER support protein folding by catalyzing disulfide bond formation, adding post-translational modifications, and promoting protein oligomerization (Cherepanova et al. 2016; Pobre-Piza and Hendershot 2021). Once proteins are properly folded, they are trafficked in vesicles to individual compartments such as the Golgi apparatus or the plasma membrane (Sun and Brodsky 2019). Despite there being a system dedicated to proper folding, many proteins fail to reach their functional state because of inefficient translocation into the ER or the failure to be post-translationally modified. These problems are exacerbated through protein misfolding, errors in translation, genetic mutations, pathological insults, age-induced damage, and the accumulation of unassembled protein subunits, referred to as orphan proteins, which prevents them from achieving functional status (Chen et al. 2011; Shao and Hegde 2016; Tao and Conn 2018).

The ER contains an elaborate protein quality control network that guarantees high-fidelity protein synthesis and folding (Adams et al. 2019). In addition, terminally misfolded proteins that do not reach their native conformation are eventually eliminated to minimize proteotoxicity. Terminally misfolded proteins are retained in the ER and targeted for degradation by the proteasome through ER-associated degradation (ERAD) and lysosomes/vacuoles through ER-phagy. ERAD is the principal degradation pathway in the ER and well conserved from yeast to mammals. It consists of targeting and ubiquitinating protein substrates by a dedicated E3 ligase. All ERAD pathways require the Cdc48 ATPase (p97 in mammals) as an energy source for substrate removal (Bays et al. 2001; Ye et al. 2005; Bodnar and Rapoport 2017; Neal et al. 2017; Twomey et al. 2019), entailing extraction of integral membrane substrates from ER membrane or movement of luminal ERAD-L substrates across the ER membrane process known as retrotranslocation followed by degradation by the cytosolic proteasome (Hampton and Sommer 2012; Lemberg and Strisovsky 2021). In yeast, ERAD is mediated by the HRD (HMG-CoA reductase degradation) and DOA (degradation of α2) pathways, both of which are conserved in all eukaryotes (Hampton et al. 1996; Laney and Hochstrasser 2003). In the HRD pathway, the E3 ubiquitin ligase HMG-CoA reductase degradation (Hrd)1 recognizes and ubiquitinates a variety of substrates, including misfolded membrane substrates (ERAD-M) and luminal substrates (ERAD-L) (Carvalho et al. 2006; Chen et al. 2006; Foresti et al. 2013; Wangeline and Hampton 2018). In the DOA pathway, the E3 ubiquitin ligase Doa10 recognizes and ubiquitinates misfolded soluble and membrane proteins, often with lesions in the cytosolic domain (ERAD-C) (Laney and Hochstrasser 2003; Ravid et al. 2006). Moreover, unassembled ER subunits escaping to the inner nuclear membrane are targeted by the Asi complex (Foresti et al. 2014; Khmelinskii et al. 2014; Natarajan et al. 2020). In addition to misfolded proteins, ERAD targets normally folded proteins. This includes the most thoroughly studied rate limiting enzyme HMG-CoA reductase, which undergoes regulated degradation as part of the cellular control for sterol biosynthesis; a process highly conserved from yeast to mammals (Hampton 2008; Jo and Debose-Boyd 2010; Elsabrouty et al. 2013). Nevertheless, some terminally misfolded proteins escape the ER, translocate through the secretory pathway, and are eventually targeted for lysosomal degradation through quality control mechanisms in the Golgi and plasma membrane or targeted for proteasome degradation by endosome Golgi-associated degradation (EGAD, discussed in the section Dsc2 and UBAC2) (Arvan et al. 2002; Schmidt et al. 2019). The ER protein quality control network is modulated by the unfolded protein response (UPR). The UPR senses ER proteotoxicity and evokes a transcriptional response to restore protein homeostasis. This is mediated through multiple mechanisms, such as the up-regulation of mediators of folding and degradation and a reduction in protein load in the ER. In addition, another defense against ER protein overload is through the ER preemptive quality control (ERpQC), which limits further protein loading into the ER during ER stress, alleviating the burdened ER (Kadowaki et al. 2018). Nevertheless, when the capacity of the UPR to maintain proteostasis is overwhelmed, cells undergo apoptosis (Travers et al. 2000; Walter and Ron 2011; Ho et al. 2020).

The rhomboid superfamily is emerging as a critical player in ER protein quality control, in which several members function in ERAD, ERpQC, and EGAD (Kandel and Neal 2020; Lemberg and Strisovsky 2021). In general, the rhomboid superfamily is present in all kingdoms of life and shares a core six-transmembrane-helix bundle that is stabilized by the highly conserved WR motif and a GxxxG motif found in loop 1 (L1) and transmembrane 6 (TM6), respectively (Fig. 1; Lemberg and Freeman 2007; Freeman 2014). Previously, Lemberg and Freeman performed an extensive analysis of gene lineages that express proteins possessing these rhomboid-like features. Based on this analysis, rhomboid proteins were placed into two general categories: rhomboid intramembrane proteases and pseudoproteases (Lemberg and Freeman 2007). Rhomboid proteases cleave membrane-anchored substrates through their serine-histidine dyad (Brooks and Lemieux 2013; Düsterhöft et al. 2017; Tichá et al. 2018), whereas rhomboid pseudoproteases do not possess a proteolytic cleavage site (Adrain and Cavadas 2020; Kandel and Neal 2020). Rhomboid pseudoproteases—derlins, UBA domain-containing (UBAC)2/Dsc2, and iRhoms—and the rhomboid protease RHBDL4, have been implicated to function in ER protein quality control (Fig. 1). Specifically, derlins, iRhoms, UBAC2, and RHBDL4 have been implicated in ERAD, whereas Dsc2 has a prominent role in EGAD.

Figure 1.

Rhomboid proteins involved in endoplasmic reticulum (ER) protein quality control. Depiction of members from the intramembrane protease and pseudoprotease subclass: RHBDL4, iRhom1&2, and Derlin-1, -2, and -3 are found in mammals, and Der1 and Dfm1 are found in yeast. All rhomboid members diverged from a common ancestral rhomboid-like core including six transmembrane helices (in blue), serine-histidine catalytic dyad, and rhomboid motifs, WR and GxxxG, and evolutionarily diverged into the intramembrane and pseudoprotease subclass. Depicted in red are additional helices or domains that are unique to each rhomboid protein. The following domains are indicated: (VBM) valosin/p97-binding motif (also known as SHP box), (UIM) ubiquitin-interacting motif, (IRHD) iRhom homology domain, (UBA) ubiquitin-associated domain.

In this review, we first describe the mechanistic and structural features of various rhomboid proteins involved in ER protein quality control, and then outline their cellular and biological roles in health and disease (see Table 1 for overview of physiological and pathological roles). Finally, we discuss future perspectives regarding the potential use of rhomboid proteins as novel therapeutic targets for mitigating the universal biological stress associated with protein misfolding.

Table 1.

Overview of rhomboids in physiology and disease

Rhomboid	Physiology	Disease
RHBDL4	Endoplasmic reticulum (ER)-associated degradation (ERAD), modulation of glycosylation	Colorectal cancer, glioblastoma, Charcot–Marie–Tooth disease, Alzheimer's disease
Derlins
Derlin-1	ERAD, preemptive quality control	Breast cancer, colon cancer
Derlin-2		Diabetes
Derlin-3		Breast cancer
UBAC2	ERAD, endosome Golgi-associated degradation (EGAD), lipid homeostasis, WNT signaling	Behçet disease
iRhoms
iRhom1	ERAD, EGFR signaling, inflammation and immune response	Breast, epithelial, colorectal cancer
iRhom2		Tylosis with esophageal cancer, rheumatoid arthritis, lupus

RHOMBOID PSEUDOPROTEASES

Derlins

Derlins are ER-resident integral membrane proteins classified as rhomboid pseudoproteases (Greenblatt et al. 2011). They play roles in a large range of processes, including ERpQC (described in more detail below) and ERAD (Knop et al. 1996; Sato and Hampton 2006; Greenblatt et al. 2011; Kadowaki et al. 2018; Neal et al. 2018). Two derlins, Der1 and Dfm1, are expressed in yeast, and three derlin pseudoproteases, Derlin-1, Derlin-2, and Derlin-3, are expressed in mammals (Knop et al. 1996; Lilley and Ploegh 2004, 2005; Sato and Hampton 2006).

Derlins’ Function in ERAD

The mechanisms associated with derlin-mediated retrotranslocation has been characterized. In vitro and in vivo assays, as well as structural analyses have revealed that the HRD complex retrotranslocates ERAD-L substrates (Baldridge and Rapoport 2016; Wu et al. 2020). The HRD complex is composed of the E3 ligase Hrd1 (containing a cytosolic RING for ubiquitination), the mostly luminal Hrd3, the luminal Yos9, Usa1 (containing a Ubiquitin-like Ubl domain), and the derlin pseudoprotease, Der1 (Carroll and Hampton 2010; Mehnert et al. 2015; Vashistha et al. 2016). Hrd3 and Usa1 are required for stabilizing the E3 ligase Hrd1 and enhancing its ubiquitination activity (Vashistha et al. 2016). During retrotranslocation, Hrd1 and Der1 form two “half channels” that are juxtaposed in a thinned membrane region. To initiate retrotranslocation, ERAD-L substrates are recruited by Hrd3 and Yos9 in the lumen, where the substrate is retrotranslocated as a hairpin loop and moved across the distorted and thinned lipid bilayer (derlin-mediated lipid thinning nature is discussed below) (Wu et al. 2020).

A phenotypic screen in yeast showed that the derlin rhomboid pseudoprotease Dfm1 has a broad role in the retrotranslocation of integral membrane ERAD-C and ERAD-M substrates (Neal et al. 2018). Previously, Dfm1 was thought to not be involved in ERAD (Sato and Hampton 2006; Goder et al. 2008), but it was later shown that the loss of Dfm1 leads to rapid suppression, masking its role in ERAD (Neal et al. 2020). During suppression when Dfm1 is absent, the HRD complex is remodeled and shifts from its normal function in ERAD-L retrotranslocation to ERAD-M retrotranslocation (Neal et al. 2020). Nevertheless, in normal circumstances, Dfm1 is a major mediator for ERAD, and Nejatfard et al. (2021) delineated the following steps for Dfm1-mediated retrotranslocation: (1) transient binding of ERAD substrate to Dfm1's L1 region, (2) Dfm1 recruitment of Cdc48 by its carboxy-terminal SHP tail, (3) Cdc48 binding to ubiquitin chain attached to ERAD substrate, (4) ERAD substrate extraction from the ER membrane, and (5) proteasomal degradation in the cytosol (Fig. 2). Notably, substrate binding through the widely conserved L1 region also occurs in its bacterial rhomboid protease ancestor GlpG, suggesting Dfm1's binding mechanism may be evolutionarily conserved (Zoll et al. 2014). A recent cryo-electron microscopy (cryo-EM) study of Dfm1's human homolog, Derlin-1, sheds light on the mechanistic features of the derlin retrotranslocation function. Rao et al. (2021) demonstrated that Derlin-1 forms a homotetrameric channel and a channel-like pore with TM1, TM2, TM5, L5, and L6, indicating that Derlin-1 possesses channel-like activity (Rao et al. 2021). Accordingly, establishing a high-resolution structure of a complex containing both a derlin and its substrate would help decipher the mechanism(s) at play for derlin-mediated retrotranslocation.

Figure 2.

Model of derlin Dfm1-mediated retrotranslocation of integral membrane substrates. (1) Dfm1 targets and binds integral membrane substrates through its L1 region. (2) Once in close proximity, the polyubiquitin chain attached to substrates binds directly to Cdc48 that was recruited by Dfm1. (3) Lipid-thinning function by Dfm1 facilitates the extraction of integral membrane substrates from the endoplasmic reticulum.

A notable feature of the rhomboid superfamily is their ability to induce lipid thinning in the surrounding membranes (Bondar et al. 2009; Kreutzberger et al. 2019). By using cryo-EM and molecular dynamic (MD) simulations, lipid thinning effects of derlins have been identified in yeast and humans, which is postulated to lower the energy costs associated with retrotranslocating membrane substrates. Nejatfard et al. (2021) and Wu et al. (2020) conducted an in-depth study of the lipid thinning effects of derlins using MD simulation and identified a cluster of hydrophilic residues, NHLST and RSSQ within TM2, that are critical for the lipid distortion function of Der1 and Dfm1, respectively. These clusters directly interacted with the phosphate head groups of the lipid bilayer (Wu et al. 2020; Nejatfard et al. 2021). For yeast Dfm1 and Der1, swapping the hydrophilic residues for hydrophobic residues within this cluster abolishes their retrotranslocation function, which implies that lipid distortion facilitates the movement of substrates across the ER membrane. Notably, sequence alignment of yeast and human derlins demonstrate an analogous cluster of hydrophilic residues exists in human derlins Derlin-1, Derlin-2, and Derlin-3, suggesting that lipid thinning is a generalized mechanism involved in the effect of all derlin rhomboid pseudoproteases (Fig. 3).

Figure 3.

T-Coffee alignment of the transmembrane regions of Derlin-1, Derlin-2, and Derlin-3 from Homo sapiens and Der1 and Dfm1 from Saccharomyces cerevisiae. Identically and similarly conserved residues are highlighted in red and yellow, respectively. The hydrophilic cluster postulated to have a role in lipid distortion is underlined in green. The following are hydrophilic amino acids (in green) within the hydrophilic cluster implicated to participate in lipid thinning: Dfm1, RSSQ; Der1, NHSLT; Derlin-1, QYSTR; and Derlin-2,3, RYCRM.

The closest human relatives to the yeast derlins Der1 and Dfm1 are Derlin-1, Derlin-2, and Derlin-3. All three human derlins are involved in retrotranslocation and the degradation of ER luminal and membrane substrates (Greenblatt et al. 2011). Human Derlin-1 is most closely related to yeast Dfm1 (20% identity) and was initially identified via its interaction with cytomegalovirus protein US11 for ERAD of MHC I (Lilley and Ploegh 2004). Biochemical studies have identified a number of ERAD substrates that are targeted by derlins. These substrates include the cystic fibrosis transmembrane conductance regulator (CFTR), disease-associated mutant CFTR-ΔF508, the epithelial sodium channel (ENaC), proinsulin, sonic hedgehog protein, the null Hong Kong mutant of α1-antitrypsin, and nonstructural (NS) viral proteins, which are critical for viral propagation (Oda et al. 2006; Huang et al. 2013; Hoelen et al. 2015; You et al. 2017; Tabata et al. 2021). Altogether, derlins act on substrates that are fundamental to a plethora of physiological processes.

Derlins’ Function in ERpQC

Aside from functioning as a major ERAD participant, Derlin-1 functions in ERpQC. During acute ER stress, ERpQC limits further protein loading into the ER, alleviating the burdened ER (Kadowaki et al. 2018). To do this, the hydrophobic signal sequences of newly synthesized proteins are recognized by the carboxy-terminal end of Derlin-1, followed by ubiquitination of a newly synthesized protein by E3 ligase Hrd1, and degradation by the cytosolic proteasome (Kadowaki et al. 2015, 2018). It will be important in the future to determine whether other derlin members function in this pathway and whether derlin mechanistic actions are the same in ERpQC versus ERAD.

Physiological Role of Derlins

Because of the lethality of the global knockout (KO) of Derlin-1 and Derlin-2 in mice, the broader physiological roles of derlins have proved difficult to study (Dougan et al. 2011; Eura et al. 2012). Knocking out Derlin-1 is lethal, whereas Derlin-2 KO mice typically die within the first 24 hours of postnatal life (Dougan et al. 2011; Eura et al. 2012). Individuals that survive this period develop skeletal dysplasia and chronic activation of the UPR (Dougan et al. 2011). Interestingly, the global KO of Derlin-3 mice is not associated with mortality, and the mice grow normally (Eura et al. 2012). Derlin-1 and Derlin-2 lethality is most likely due to their ubiquitous expression in all tissues, whereas Derlin-3 is predominantly expressed in the placenta, pancreas, spleen, and small intestine (Eura et al. 2012). Moreover, the normal growth of Derlin-3 KO mice may be due to Derlin-1 and Derlin-2 compensation. This is not surprising given many proteins within the ER protein quality network serve redundant functions.

This lethality demonstrated not only shows that derlins are physiologically important, but also stresses the need for generating tissue or cell-specific KO studies. Indeed, numerous cell-specific derlin KO mice have been generated, and these have offered new insights into their physiological significance. For example, Schwann cell-specific Derlin-2 KO mice develop neuropathy because of an impairment in ERAD within peripheral nerves and defects in myelin morphology and function (Volpi et al. 2019). Most recently, Derlin-1 or Derlin-2 KO in the central nervous system (CNS) of mice caused volume loss and atrophy of the cerebellum and striatum, leading to impaired motor function (Sugiyama et al. 2021). Unexpectedly, Derlin-1 or Derlin-2 CNS-specific KOs, exhibited reduced mRNA levels of genes associated with cholesterol biosynthesis, resulting in a total decrease of cholesterol levels (Sugiyama et al. 2021). This seminal finding is the first evidence that derlins are involved in regulating cholesterol biosynthesis. Indeed, both Derlin-1 and Derlin-2 are required for SREBP-2 activation and the transcription of genes involved in cholesterol synthesis, which is critical for neurite outgrowth and postnatal development (Sugiyama et al. 2021). This study establishes a new physiological role for derlins in cholesterol regulation and brain development. Future studies on derlins’ mechanistic actions in mediating SREBP-2 action will provide invaluable insights on derlins’ regulatory role in cholesterol metabolism.

Hepatocyte-, podocyte-, and B-cell-specific Derlin-2 KO have also been generated (Dougan et al. 2011; Ren et al. 2018), and while all these cell types typically retain their function, most, if not all, show dysregulation of ERAD and chronic ER stress (Dougan et al. 2011; Ren et al. 2018). These mice develop kidney diseases, such as diabetic nephropathy, renal fibrosis, and ischemia-reperfusion injury, which are likely to be long-term consequences of this ER stress (Ren et al. 2018). Increasing the levels of Derlin-2 improves survival of podocytes with ER stress, probably because ER protein imbalance is corrected (Ren et al. 2018). Patients with diabetic nephropathy who generate high expression of Derlin-2 may be able to combat the chronic ER stress (Ren et al. 2018). Overall, derlins have far-reaching roles in vivo that are underpinned by the management of ER stress. Accordingly, defects in derlins are likely to have physiological consequences, as demonstrated by the numerous phenotypes of derlin mutant mouse models.

Derlins’ Role in Cancer Progression

ER stress is a feature of numerous diseases, including cancers (Corazzari et al. 2017; Moon et al. 2018), because of the high cell proliferation rate and poor control over protein synthesis, correct folding, and trafficking (Clarke et al. 2014). It is perhaps unsurprising that derlins play an important role in several cancers, given their important roles in the control of ER stress via ERAD. Derlin-1 and Derlin-3 are both up-regulated in breast cancer (Wang et al. 2008; Shibata et al. 2017). In colon cancer, Derlin-1 is highly expressed (Tan et al. 2015), and Derlin-3 expression is repressed because of promoter hypermethylation (Lopez-Serra et al. 2014). The latter finding is perhaps counterintuitive, but the hypermethylation may act as a metabolic switch to enable the stabilization of GLUT1, a glucose transporter, thereby increasing cellular energy supply (Lopez-Serra et al. 2014). Typically, GLUT1 is targeted for degradation by Derlin-3, and therefore the stabilization of GLUT1 induced by the down-regulation of Derlin-3 increases the energy supply to cells, supporting their proliferation (Lopez-Serra et al. 2014). Due to cancer cells’ susceptibility to ER stress, derlins are important therapeutic targets for ameliorating or delaying cancer disease.

Dsc2 and UBAC2

Like derlins, yeast Dsc2 and its mammalian homolog UBAC2 are ER-resident rhomboid pseudoproteases that participate in EGAD and ERAD, respectively (Yamazoe et al. 2017; Schmidt et al. 2019).

Dsc2's Function in EGAD

Dsc2 is a key component of the DSC (defective for SREBP cleavage) complex, which is composed of the E3 ligase Dsc1, the rhomboid pseudoprotease Dsc2, and the adaptor proteins Dsc3 and Dsc4 (Lloyd et al. 2013; Hwang et al. 2016; Schmidt et al. 2019). Dsc2 possesses six transmembrane helices and a carboxy-terminal cytosolic UBA domain (Fig. 1). In general, UBA domains form three-helix bundles with low sequence conservation have been shown to bind to Lys-63 di-ubiquitin, but not Lys-48 di-ubiquitin in vitro, indicating that Dsc2 functions in a post-ubiquitination step. Current knowledge on DSC function demonstrates that they function in lipid homeostasis. In eukaryotes, cholesterol concentrations are principally controlled by the cleavage and activation of SREBP (Shao and Espenshade 2012; Shimano and Sato 2017). A body of study has delineated the mechanism of the DSC-mediated activation and cleavage of the fission yeast homolog of SREBP, Sre1 (Stewart et al. 2012; Lloyd et al. 2013; Kim et al. 2015; Hwang et al. 2016). In general, when sterol concentrations are high, SREBP is inactive, and therefore sterol uptake and synthesis are low. However, when sterol concentrations are low, SREBP is cleaved by the DSC complex, then translocated to the nucleus, where it up-regulates the transcription of genes involved in sterol influx and biosynthesis (Stewart et al. 2012; Hwang et al. 2016; Schmidt et al. 2019). Beyond the regulation of sterols, other roles of the DSC ubiquitin complex have been characterized in baker's yeast, in which it targets Orm2 for degradation via the EGAD pathway (Schmidt et al. 2019). Importantly, Orm2 is a negative regulator of sphingolipid biosynthesis (Breslow et al. 2010; Han et al. 2010; Liu et al. 2012), which is vital in numerous cellular processes, including cell growth, migration, apoptosis, and inflammation (Hannich et al. 2011; Hannun and Obeid 2018). In the future, it will be interesting to determine whether the rhomboid-like features of Dsc2 are required for Sre1 activation and Orm2 degradation.

UBAC2's Function in ERAD

Mammalian UBAC2, like Dsc2, contains a UBA domain that is able to bind polyubiquitin chains, and it is therefore likely to be involved in ERAD (Christianson et al. 2012). Consistent with this, the depletion of UBAC2 results in the stabilization of α1-antitrypsin, a canonical ERAD substrate (Christianson et al. 2012). In addition, UBAC2 regulates UBXD8, which has roles in both ERAD and lipid droplets (LDs) (Olzmann et al. 2013). UBAC2 restricts the trafficking of UBXD8 from the ER to LDs, where it acts as a rate-limiting enzyme in lipid hydrolysis (Olzmann et al. 2013). Therefore, UBAC2 influences cellular fat storage and plays a role in energy homeostasis.

Like derlins, UBAC2 delivers substrates to the E3 ligase gp78 for ubiquitination. Both UBAC2 and gp78 can form a complex with substrate adaptor Limb region 1 protein homolog-like (LMBR1L), which can recruit Frizzled-6 for ubiquitination and degradation (Choi et al. 2019). Importantly, Frizzled-6 is a transducer of WNT signaling, a pathway critical for embryogenesis and oncogenesis (Choi et al. 2019). Thus, Dsc2 and UBAC2 are important in ERAD and EGAD, but there is much still to be learned regarding the substrates targeted by Dsc2 and UBAC2 and their roles in development and physiology.

UBAC2's Role in Behçet Disease

Currently, there is limited information on the role of UBAC2 in disease. Single nucleotide polymorphisms (SNPs) have been identified in genome-wide association studies (GWAS) of UBAC2 and these SNPs are strongly associated with Behçet disease (Yamazoe et al. 2017). Behçet disease is a rare chronic inflammatory disease that is characterized by lesions and inflammation throughout the body, but typically found in the eye and CNS (Yamazoe et al. 2017). UBAC2 SNP levels are also elevated in Behçet disease; however, the reason for this is still unknown (Yamazoe et al. 2017). Notably, other protein degradation machineries such as UBASH3B, SUMO4, and UBEQL1 are associated with Behçet disease suggesting that the ubiquitination pathway plays an important role in its pathology (Yamazoe et al. 2017). More research is required to examine the underlying mechanisms associated with UBAC2 function in the context of this disease.

iRhoms

iRhoms are rhomboid pseudoproteases that play prominent roles in ERAD and intracellular trafficking from the ER to Golgi. They are more closely related to rhomboid proteases than to derlins. They were first identified and characterized in Drosophila, which expresses a single iRhom, whereas mammals express two: iRhom1 and iRhom2 (for review, see Dulloo et al. 2019).

iRhoms contain seven transmembrane helices and an extended cytoplasmic amino terminus (Fig. 1). A hallmark feature is the presence of a highly conserved Cys-rich luminal loop 1, called the iRhom homology domain (IRHD) (Lee et al. 2016). Another distinct characteristic of iRhoms is the presence of a conserved proline residue adjacent to the expected catalytic serine of the proteolytic site (i.e., GPx replaces the GxS rhomboid catalytic motif) (Lemberg and Freeman 2007; Bergbold and Lemberg 2013).

iRhom's Role in ERAD

The relationship between iRhoms and ER protein quality control was first established by Zettl et al. (2011) who showed that the sole iRhom expressed in Drosophila promotes ERAD, and that iRhom in the ER negatively regulates EGFR signaling by promoting the ERAD of EGF ligands. The results of this seminal study provide a model for understanding the precise regulation of signaling pathways. In Drosophila, iRhom is expressed in the brain and CNS, and iRhom mutants spend long periods of time in a sleep-like state (Zettl et al. 2011). A similar sleep-like phenotype is caused by an up-regulation of EGFR signaling, which is consistent with iRhom negatively regulating EGFR activity and promoting wakefulness (Zettl et al. 2011). Using genetic and cell biology approaches, the authors demonstrated that the inhibition of EGFR signaling is suppressed by the knockdown of E3 ligase Hrd1, which further establishes a contribution of ERAD to the regulation of EGFR signaling (Zettl et al. 2011). Contrary to the effect of Drosophila iRhom in inhibiting EGFR signaling, the two mammalian iRhoms, iRhom1 and iRhom2, induce EGFR signaling by promoting the trafficking and activation of the sheddase protease ADAM17 (also known as TACE) (Christova et al. 2013; Li et al. 2015).

Studies of viral infection in mammals have also shown a role for iRhoms in ERAD. Unlike the effect of Drosophila iRhom on promotion of ERAD, mammalian iRhoms prevent the ERAD of two antiviral response factors, STING and VISA (also known as MAVS, IPS-1, and Cardif) (Luo et al. 2016, 2017). STING responds to DNA virus infections by moving from the ER to the perinuclear microsome, where it induces type 1 interferon expression (Luo et al. 2016). iRhom2 recruits the deubiquitinating enzyme EIF3S5 to STING, which rescues it from proteasomal degradation (Luo et al. 2016). In addition, iRhom2 mediates the auto-ubiquitination and degradation of the E3 ligase RNF5 and disrupts the ERAD of VISA, such that it can perform its essential role in the innate immune response to RNA viruses. Furthermore, during the later stages of RNA virus infection, iRhom2 promotes the degradation of the E3 ligase MARCH5 and inhibits the mitochondrion-associated degradation of VISA (Luo et al. 2017). These studies are in agreement with mice studies, which demonstrate that the loss of iRhom2 leads to increased susceptibility of mice to Listeria monocytogenes infections, typically associated with foodborne illness (McIlwain et al. 2012). Moreover, the deficiency of iRhom2 in mice leads to the inability to induce antiviral genes in response to viral infections such as herpes simplex virus type 1 (HSV-1), Sendai virus and vesicular stomatitis virus (VSV) (Luo et al. 2016). Overall, iRhoms have far reaching clinical implications because of their regulatory role in innate immunity.

iRhom’s Role in Cancer

A growing body of literature suggests a pathological role for iRhoms in cancer (Lemberg and Adrain 2016; Adrain and Cavadas 2020). In particular, iRhom1 expression is up-regulated in breast, epithelial, and colorectal cancers, and has been shown to be associated with increased rates of disease progression and metastasis (Yan et al. 2008; Hosur et al. 2018). Furthermore, this relationship has important clinical consequences, because high iRhom1 expression is linked to a poor response to chemotherapy, and thereby poor survival (Yuan et al. 2018). Moreover, iRhom1 deletion in breast cancer cell lines leads to apoptosis and reduced tumor growth in mice xenografts, implicating iRhoms as potential therapeutic targets for treating cancer (Yan et al. 2008).

RHOMBOID PROTEASE

RHBDL4

Mammalian RHBDL4 is classified as a rhomboid protease and has been best characterized for its role in ERAD (Fleig et al. 2012). RHBDL4 consists of six transmembrane helices, with a serine-histidine catalytic dyad within the membrane plane between TM4 and TM6 (Fig. 1). It also contains a p97/valosin-binding motif (VBM) and ubiquitin-interacting motif (UIM) at its carboxyl terminus, which suggests that the molecular function of RHBDL4 involves coordinated ubiquitin binding and p97/VCP recruitment (Fleig et al. 2012; Lim et al. 2016).

RHBDL4's Function in ERAD

Numerous ERAD substrates are targeted by RHBDL4, being cleaved into fragments and retrotranslocated to the cytosol for proteasomal degradation. These cleavage events can occur within the transmembrane helix or the ectodomain of the substrate (Fleig et al. 2012; Recinto et al. 2018). Moreover, RHBDL4 targets not only membrane substrates, but also luminal substrates, by cooperating with the adaptor proteins Erlin-1 and Erlin-2 (Kühnle et al. 2019). Thus, RHBDL4 targets a vast range of substrates. A detailed structural analysis by Lim et al. (2016), elucidated the sequence of events associated with RHBDL4-mediated cleavage and the retrotranslocation of transmembrane substrates during ERAD. The authors proposed that aberrant positively charged transmembrane substrates are targeted by RHBDL4 by the direct binding of destabilized transmembrane residues to the proteolytic site and subsequent binding of the polyubiquitin chain of substrates by the UIM. Following intramembrane proteolysis, p97/VCP is recruited by binding to the VBM of RHBDL4, and because ubiquitin has a higher affinity for p97/VCP than UIM, RHBDL4 transfers the ubiquitinated cleaved substrate to p97 for cytosolic proteasomal degradation (Lim et al. 2016). This is intriguing because substrate cleavage and release involves a coordination of events and information from both sides of the ER membrane.

RHBDL4 Targets Substrates Involved in Neuronal Maintenance and Cholesterol Homeostasis

ERAD substrates are numerous and diverse. Lemberg's group sought to identify substrates targeted by RHBDL4-mediated ERAD by stable isotope labeling using amino acids in cell culture (SILAC) and identified proteins of the OST complex that are responsible for the glycosylation of newly synthesized proteins (Knopf et al. 2020). Therefore, RHBDL4 can regulate the glycosylation of proteins, which is vital for numerous cellular processes, including the regulation of the immune system, cell adhesion, and ER stress reduction (Ohtsubo and Marth 2006; Knopf et al. 2020).

RHBDL4 also targets and cleaves amyloid precursor protein (APP), which prevents the conversion of APP to Aβ peptides that are prone to aggregation and associated with Alzheimer's disease (AD discussed below) (Paschkowsky et al. 2018; Recinto et al. 2018). In addition, RHBDL4 targets the α subunit of pre-T-cell receptor (pTα) and disease-associated myelin protein zero (MPZ) (discussed in the section RHBDL4's Role in Cancer and Neurological Diseases) (Fleig et al. 2012). To date, many of the RHBDL4 substrates have been studied in tissue culture. Therefore, whether some of these substrates represent bona fide endogenous ERAD substrates remains to be confirmed, but this would provide a new perspective on the biological function of RHBDL4.

RHBDL4's Role in Cancer and Neurological Diseases

The roles of RHBDL4 in disease have mostly been studied in the context of cancer. RHBDL4 expression is up-regulated in glioblastoma, colorectal, and liver cancer cell lines, and is linked to tumor growth (Wei et al. 2014; Song et al. 2015; Miao et al. 2017). Furthermore, the study of clinical samples has suggested that patients with low RHBDL4 expression in colorectal cancer tissue have better survival outcomes than those with higher RHBDL4 expression (Song et al. 2015). There are various mechanisms that might underpin the associations of RHBDL4 with cancer progression. For example, transforming growth factor α (TGF-α) cleavage and activation by RHBDL4 up-regulates EGFR signaling (Miao et al. 2017). However, this mechanism has been disputed and an alternative suggested, in which RHBDL4 protease activity promotes pro-TGF-α secretion from exosomes (Wunderle et al. 2016). Furthermore, RHBDL4 cleaves a subunit of the OST complex, which might lead to the degradation of BIK, promoting apoptosis (Knopf et al. 2020). Although it is known that RHBDL4 reduces apoptosis and promotes the proliferation of cancer cells, the mechanism whereby RHBDL4 affects cancer progression requires further investigation.

RHBDL4 is also associated with neurological diseases. The RHBDL4 substrate mutant MPZ (L170R) has been shown to be associated with the neuropathy, Charcot–Marie–Tooth disease, which is characterized by muscular weakness, wasting, and sensory loss (Fleig et al. 2012). MPZ is essential for the myelination of peripheral neurons and aids in the transmission of nerve impulses (Mandich et al. 2009; McMillan et al. 2010). Little is known about the mechanism of RHBDL4 targeting and binding of MPZ L170R for degradation and whether other ERAD components are required for this process. Additionally, there are over 120 disease-causing MPZ mutants and whether or not RHBDL4 targets and cleaves these other mutants remains to be investigated (Mandich et al. 2009). In addition, RHBDL4 is associated with AD (Paschkowsky et al. 2016, 2018). The cleavage of amyloid processing protein (APP) to form β-amyloid (Aβ) is one of the key characteristics of AD (Knopman et al. 2021). APP cleavage is mainly mediated by BACE (also called Asp2 and memapsin2) and γ-secretase proteases to produce Aβ peptides (De Jager et al. 2014; Knopman et al. 2021). Interestingly, RHBDL4 has recently been shown to cleave APP in its ectodomain, which reduces the production of the Aβ peptides (Paschkowsky et al. 2016). These results underscore the unexpected role of RHBDL4 in APP cleavage and a potential therapeutic target for treating AD. Moreover, most of the findings discussed above have been done in tissue culture; hence, much remains to be done to validate findings in in vivo models with relevance to neurological disease.

CONCLUDING REMARKS

In the context of ER protein quality control, rhomboid proteins function in protein turnover via ERAD, EGAD, and ERpQC. Accordingly, rhomboid proteins in ER protein quality control appear in vast biological roles, including development, energy balance, sterol regulation, neuronal function, inflammation, immunity, and protein quality control (Kandel and Neal 2020; Lemberg and Strisovsky 2021). Because they drive a variety of biological processes, the rhomboid superfamily underlies many of the most pressing human maladies such as Alzheimer's, immune disorders, cancer, and infectious diseases (Bergbold and Lemberg 2013). The breadth and function of rhomboid proteins shows their importance in health and diseases. Therefore, rhomboid proteins serve as potential therapeutic targets for various diseases. A plethora of biophysical and structural information of several rhomboid proteins already exists, enabling the recent development of drug inhibitors that modulate their activity (Tichá et al. 2018). However, the underlying molecular mechanism of many other rhomboid proteins is still unknown. Structural and mechanistic information gleaned from these less characterized rhomboid proteins will be informative and drive the discovery of new drugs targeting rhomboid proteins. Overall, we expect drugs targeting rhomboid proteins will treat diseases related to abnormal trafficking, ER proteotoxicity, and lipid dysregulation.